Speech recognition relates to the transcription of words spoken by a human. In performing speech recognition, acoustic and language models are used by automated speech recognition (ASR) engines to statistically analyze an encoded utterance in order to generate one or more likely transcriptions that reflect the words spoken by the speaker.